Recreational Participation among Ethnic Minorities and Immigrants In and the

Amanda Aizlewood∗ Pieter Bevelander# Ravi Pendakur∗ Social Development Canada IMER, Malmö University Social Development Canada Amanda.Aizlewood@sdc- [email protected] [email protected] dsc.gc.ca

# The author would gratefully acknowledge the generous support provided by the Swedish Foundation for International * The opinions expressed in this Cooperation in Research and report are those of the author and/or Higher Education (STINT), the contributors and do not necessarily Institute for Sociological and reflect the views of Social Economic Research (ISEO), Development Canada or the Erasmus University which allowed . us to do the research.

2

Abstract In this paper we compare the community engagement of minorities as measured by group recreational activity in Canada and the Netherlands – two countries with high immigration intake and high levels of ethnocultural diversity. Using logistic regression we ask questions concerning the determinants of recreational participation, focussing on the degree to which differences are a product of minority status or of more general socioeconomic status. We also compare rates of participation in the two countries to see if there is an underlying difference which can be explained by socioeconomic and demographic status. We find that socio-demographic characteristics are generally much stronger predictors of participation than characteristics associated with minority status, regardless of country.

3

1. Introduction

As with other forms of active participation, recreational participation involves engagement and exchange among and between individuals, families and groups. Researchers argue that such engagement is beneficial because it leads to a greater sense of mutual obligation among individuals and toward the larger community (Putnam, 1993; Brehm and Rahn, 1997; Putnam, 1995; Fukuyama, 1995). Higher participation leads to higher interaction, which in turn leads to higher trust among individuals and between groups. In Canada and the Netherlands – two countries with high immigration intake and high levels of ethnocultural diversity – it is in everyone’s interest that immigrants and ethnocultural minorities be actively and positively engaged in the civic, social and cultural spheres of their communities. The engagement of immigrants and ethnocultural minorities in recreational activity is of particular interest to social researchers because it is an indicator of community engagement. Such participation, scholars argue, can lead to social benefits, such as increased interaction across diverse groups and an increased sense of belonging to one’s community.

The purposes of this study are twofold. First, using descriptive tables we analyse the rates at which immigrants and ethnocultural minorities in Canada and the Netherlands participate in specific forms of recreational activities as compared to the native-born and majority portion of the population.1 Second, using logistic regression we investigate the factors which influence the rate of such participation and the degree to which these factors differ between Canada and the Netherlands. In particular we investigate whether ethnocultural or immigrant-status characteristics predict the likelihood of recreational participation better than sociodemographic characteristics.

We find that although there are some differences in the participation rates of ethnocultural minorities and immigrants in both Canada and the Netherlands, the dominant drivers are related to socio-economic and demographic characteristics. These drivers are consistent in both countries, with variables such as age, education and employment status having about the same magnitude and significance.

1 ‘Ethnocultural minority’ is defined for our purposes as those individuals are not members of the ethnocultural majority group as it is generally understood in the country of study. In Canada, this refers to non- British/French/Canadian ethnic origin. In the Netherlands, it refers to non-Dutch ethnic origin. It is important to note that ethnocultural identification exists separately from ‘immigrant’ status. Ethnocultural minorities can be native-born, and immigrants can be majority or non-majority in ethnic origin. For this reason, it is important to study both groups in an analysis of this kind. 4

2. The Literature

The following section provides a brief review of the link between participation, well-being and minority status, as well as a look at the immigration regimes in Canada and the Netherlands. The review begins by defining recreational participation. We then provide a brief look at theories surrounding the relationship between ethnicity and recreational participation. We review findings of previous studies that look at the impact between ethnicity, immigrant status and recreational participation, and provide highlights of findings related to the influence of sociodemographic characteristics on recreational participation.

2.1 Recreational participation defined

Recreational participation is defined as an individual’s activities during his or her leisure time. These activities can take place in either formal or informal settings (see Helly, 1997). Informal recreational participation includes those activities which take place in groups on an ad hoc basis as well as those that take place at stable, pre-arranged times. They are relatively permanent, but are not governed by institutions. Examples of this informal kind of social participation are youth groups, music groups, cooking clubs, quilting bees, and friends who meet on weekend afternoons to play a sport in a local park. Formal recreational participation includes any affiliation with registered associations, clubs and groups with an ongoing, scheduled program of activities, a known and stable meeting place, and officers with known responsibilities. Examples of this more formal kind of participation are choral societies, sports leagues, and exercise classes.

2.2 Recreational participation of ethnic minorities

Over the last two decades, a number of studies have been conducted on the recreational behaviour of ethnocultural minorities. These studies have generally concluded that ethnic background influences a variety of phenomena related to sport, leisure and recreation behaviour (see Coakley, 2001; Henderson and Ainsworth, 2001; Hutchison, 1987; Juniu, 2000; Tirone and Shaw, 1997) and that by and large, recreational participation is lower among ethnocultural minorities as compared to the majority population.

5

There are several theoretical approaches that attempt to explain this negative relationship between recreational participation and minority ethnicity. The explanations proffered for this discrepancy in participation rates have emerged largely from American scholars and are, not surprisingly, strongly influenced by a racialised class-underclass paradigm of social relations. In particular, the approach known as marginality posits that recreational participation among ethnic groups is constrained by the forces of social inequality and discrimination (see Yancey and Snell, 1976). This approach grew out of studies attempting to account for the under-representation of African in recreational activities (Floyd, 1998). It explains lower participation rates as the result of social barriers that act as obstacles to participation. The marginality hypothesis has been subject to a high degree of criticism (see Hutchison, 1988; Washburne, 1978) on the basis that it is has been accepted too uncritically and is far too simplistic in its attempt to group all as members of an underprivileged and monolithic class structure (see Woodard, 1988; Johnson, Horan and Pepper, 1997). It is also questionable as to how transportable these ideas are to Canada and where the minority populations have very different histories and where race based attitudes are not as dominant.

Similar to the marginality approach is the opportunity approach. Here, non-participation of minority ethnocultural groups in ‘mainstream activities’ is explained by a lack of sufficient access and opportunity to do so (Lindsay and Ogle, 1972). This may be due to lack of available infrastructure, such as a lack of green space, sports facilities or meeting halls. The hypothesis holds that as costs increase, opportunities for marginalized groups decline. Since ethnocultural minorities are presumed to occupy the lower rungs of the social ladder and are often spatially disadvantaged, they lack the means to access opportunities for recreational participation.

Other theoretical approaches centre on the essential ‘ethnic-ness’ of minority groups themselves to explain disparities in participation rates. The ethnicity approach argues that differences in recreational participation among ethnic groups can be explained by the existence of a distinct set of subcultural norms and values with respect to leisure activity. In this view, recreational participation among ethnic groups is the result of specific group interests and is created and directed to meet these needs (Washburne, 1978; Walter, Brown and Grabb, 1991.) Identity theory suggests that members of ethnic groups deliberately choose to participate within their own communities and not in so-called ‘mainstream’ activities because they wish to preserve their ethnic identity and heritage. Individuals consciously choose to construct and preserve their identity by engaging in ethno-specific recreation (Karlis and Dawson, 1995). In the complementary approach, ethnic groups, upon recognising that the 6 host culture cannot meet their sociocultural needs, create their own associations and organizations to fill these needs. This framework is most predominant within multicultural societies because it allows for individuals to remain active in the host culture while continuing to identify with their ethnic heritage (Karlis and Dawson, 1995).

2.3 Recreational participation of immigrants

Research in recreational participation rates suggests that immigrants participate less often than non-immigrants. However, little effort has been devoted to explaining why they participate less or how the rates of participation change over time despite the fact that the lifestyles of immigrants undergo significant shifts following the migration process. Research thus far suggests that while immigrants are subject to the same constraints to participation in recreational pursuits commonly encountered by the general population, they are also subject to a number of barriers particular to their minority status and to problems with their adjustment to a new environment (see Stodolska, 2000). The constraints that are particularly acute for immigrants, are, time crunch, problems with language, being unfamiliar with the ways of life in the host country, insufficient access to known and desired forms of recreational activity, and experiences with discrimination. Immigrants may also lack the advantages associated with early socialization in the host country, and may lack sufficient social networks and knowledge of available recreational opportunities (Stodolska, 1998).

Time constraints are particularly problematic for newcomers and can have a significant effect on their opportunities for recreational participation. Stodolska and Alexandris (2004) find that regardless of an immigrant’s ethnocultural or sociodemographic background, the initial period after settlement in the host country is usually associated with low levels of voluntary physical activity because for new immigrants, “sport and physical recreation are typically quite low on the priority list of immigrants who struggle to adjust to a new environment, who often hold several low wage, but physically demanding jobs, and who have hardly any free time available” (392-93).

From this perspective, newcomers are seen to settle in neighbourhoods where it is easier to establish social networks and maintain their cultural identity. This spatial concentration of immigrants in certain neighbourhoods, especially according to country or region of origin has implications for recreational participation. On the one hand, spatial concentration facilitates the creation of social networks, and therefore social capital, among newcomers of the same origin. On the 7 other hand, residential segregation caused by spatial concentration may limit the potential for immigrants to participate fully in the broader society (see Ley and Germain, 2000).

Many immigrants face a form of social isolation, due in part, to the opportunities for participation open to them. Breton (1997) suggests that “with immigrants, it is inevitable that social participation will occur within the ethnic community, at least at the outset. This is where social capital is most readily available to the immigrant. Belonging to the ethnic community is spontaneous and natural for the immigrant, and necessary for survival.” Tyrone and Shaw (1997) found that for immigrant women, the centrality of family and a sense of a lack of entitlement to leisure time heavily influenced lower rates of recreational behaviour. Rublee and Shaw (1991) examined the experience of Latin American refugees in Atlantic Canada and found that a lack of community participation resulted from language difficulties and reduced opportunities for socialization in church and neighbourhood settings. In their study, they found that this lack of participation could be explained by a set of immigrant-specific constraints that included post-arrival isolation, cultural differences, inadequate language skills, a lack of overall orientation in Canadian day-to-day life, and difficulties in obtaining access to affordable and culturally sensitive childcare.

Similar results have been found in the Netherlands. Surinamese and Antilleans had participation rates that were lower than native Dutch but higher than Turks and Moroccans (De Haan & Breedveld, 2000). Differences are found by generation, with higher participation rates in sport activities for the second. In a multivariate analysis, controlling for education, occupation, sex and age differences in sport activity between groups did not disappear. Differences in membership in recreational associations, however, disappear when including these explanatory variables (De Haan & Breedveld, 2000). Participation in clubs and volunteer work is clearly less frequent for immigrants as compared to the native Dutch population (de Hart, 2002). The same holds for political participation (Fennema et al., 2000). In a study using data from 1984, it was shown that Turks and Moroccans were less active in sports compared to the native Dutch (Lagendijk & Van der Gugten, 1995).

Recreational participation can also be positively affected by immigration. Studying the behaviour of recent immigrants in , Stodolska (2000) found that positive changes in post- arrival participation rates could be explained in part by the decreased impact of behavioural constraints, and the increased access and exposure to new recreational opportunities. Using Jackson and Dunn’s 1988 model, she argued that the life changes that may be attributed to the immigration experience can be expected to make immigrants more likely to cease participation in some of their old 8 recreational activities. At the same time, however, immigrants will not only exclude old activities from their leisure repertoire but also replace their old pastimes with ones learned in the host country.

2.4 Sociodemographic influences on recreational participation

In addition to the impacts of ethnocultural minority and immigrant status, other personal, sociodemographic characteristics influence the rate at which individuals and groups participate in recreational activities.

The relationship between urbanization, ethnic diversity, and participation seems particularly relevant to Canada and the Netherlands. Both are highly urbanized societies with immigrant settlement concentrated in a few urban centres. In Canada, almost nine out of ten immigrants live in the major . Sixty-two percent of immigrants live in , or . In the Netherlands, immigrants are concentrated in the largest cities, particularly the four largest (, , Rotterdam and Utrecht) where almost half of all ethnic minorities live. In both countries, recent immigration is dominated by intake from Africa, the Caribbean and Asia, rather than from European sources (Pendakur, et al, 2003; Uunk, 2002; Martinez & Vreeswijk, 2002). Urban living occupies a central role in studies of social contact because of the implication of life in the decline of traditional forms of community-based interaction (see Guest and Wierzbicki, 1999). Urbanization is seen as the producer of “a social order in which the traditional ties of community- shared space, close kinship links, shared religious and moral values were being replaced by anonymity, individualism and competition.” (Forrest and Kearns, 2001). Alesina and Ferrara (1999: 2) find that after controlling for individual characteristics, scores on participation measures are significantly lower in more ethnically fragmented communities, the vast majority of which are located in urban settings.

Educational attainment also influences rates of participation. Helliwell and Putnam (1999) build on the work of Nie, Junn, and Stehlik-Barry (1996) by investigating the impact of education on participation. They find that with respect to organization memberships, an individual’s educational attainment has a positive effect on level of membership. They explain the result using ideas about the role of social trust in creating good conditions for participation, arguing that “higher than average education levels may help to create a climate of trust that is self-reinforcing. If individuals know that higher education levels make others more likely to be trusting … then they are more likely to trust others.” Alesina and La Ferrara (1999:20) find that in areas of high cultural diversity, years of 9 schooling are positively associated with participation. These correlations remain highly significant and stable in the face of multiple controls both in Canada and the Netherlands (see SCP, 2000).

Women and men participate at different rates. Alesina and La Ferrara (1999: 20) find that in neighbourhoods with high levels of ethnocultural heterogeneity, women participate significantly less than men. Research from Britain and Belgium on the ethnic impact on recreational participation concentrated on the recreational experiences of South Asian girls (see Verma and Darby, 1994; de Knop et al., 1994). These studies found that South Asian girls were significantly constrained in their recreational participation, particularly in those activities taking place outside the home and in sports participation. The reasons for this constraint were a lack of parental approval for these activities, parental enforcement of strict dress codes, an inadequate availability of single-sex facilities, and their own religious beliefs about the proper behaviour of females. Relative to their female counterparts, South Asian boys experienced greater freedom but were more likely to experience racial discrimination. Duyvendak et al. (1998) studied the participation in sports clubs in the city of Rotterdam and showed that lower rates for ethnic minorities like the Surinamese, Moroccans and Turks compared to the native Dutch were primarily explained by different levels of participation by women of these minorities relative to native Dutch women. Moroccan women in particular had very low participation rates. On the other hand, Surinamese men had a higher rate of participation than native-born Dutch men.

2.5 Immigration in Canada and the Netherlands

Overall, there is general consensus that ethnocultural minorities and immigrants have lower levels of participation than the native-born population. Our question concerns the degree to which this is true in Canada and the Netherlands. We ask this question because until recently, both countries had similarly open immigration policies and broadly similar intake patterns.2 However, there is a crucial difference because immigrants to Canada are viewed as permanent migrants whereas this is less likely to be the case in the Netherlands. Where Canada can be viewed as a pro-immigration country that uses a conscious but selective immigration policy and sees immigrants as new citizens, the Netherlands is in many ways a immigration country ‘against its will’ (Penninx & Schrover, 2001). There are however, a number of broad similarities. Both countries have witnessed massive transformations in the type of intake especially since the Second World War; both have had relatively

2 While there has been a retreat in recent years from multiculturalist policies in the Netherlands, this was not the case at the time that the survey was conducted in that country. 10 open immigration policies and both are officially multicultural societies which at present offer relatively open naturalization.

In the Netherlands, in-migration occurred first through colonial and labour migration. In the mid 1970s, family reunion and marriage migration dominated, and by the 1980s and early 1990s asylum seekers became far more prevalent (Doomernik, 2005). In 2000, approximately 66 percent of the received migrants and their descendents in the Netherlands were from the Mediterranean (Turkey and ) and the Caribbean (Suriname and Dutch Antilles) region (Groeneveld & Weijers-Martens, 2003). Intake from these countries alone contributed more than one million people to the country. Immigration intake to Canada prior to 1960 was primarily from European sources and was based on sponsorship and family reunification. New regulations introduced in the early 1960s added a labour force component to selection. Over time, regulations concerning regionally based intake were slowly removed, creating for the first time in Canada an arguably ‘colour free’ intake policy. As a result, the dominant source countries slowly shifted away from Europe and toward Asia and the Caribbean. Today, about 13 percent of have non-European origins.3 (Pendakur, 2000)

There are some fundamental differences in how immigration is viewed in Canada and the Netherlands. Canadian policy has always viewed immigration to be comprised of permanent migrants. Immigrants are expected to come to Canada, settle and contribute to society. This is less the case in the Netherlands. Granted, it is more or less accepted that migrants who have been in the country will stay permanently, but permanent migration is not a policy goal in itself. Further, new migration tends to be viewed as something to be discouraged (Penninx & Schrover, 2001). Despite this, (as is the case in Canada), the Netherlands is classified as a country with a multicultural policy which embodies a concept which could be described as ‘integration with maintenance of own culture’. In the Netherlands the original intent of the policies was more closely aimed at non-western ethnic minorities than is the case in Canada. On the socio-economic level, the main goal of it was the improvement of the position of disadvantaged ethnic minorities.

At the political level, naturalization is relatively straightforward and is possible after three years of residence in Canada and five years of residence in the Netherlands. Further, both Canada and the Netherlands have employment equity/affirmative action type programs designed to elevate the socio- economic position of poorly represented groups. At the cultural level, the Netherlands appears to have more far reaching programs than is the case for Canada, particularly in the area of recreation and

3 This figure does not include the 3% of people who identify as Aboriginal persons. 11 participation. The policy encourages an active role by ethnic organizations through subsidization. It was this policy that could be the reason for the increase in the number of minority organizations as well as their goals and orientation. The integration policy has been questioned rather heavily during the years since the 9/11 incident. This combines with the subsequent rise of populist right-wing politicians has given rise to a more assimilationist policy (Doomernik, 2005).

The preceding review suggests that Canada and the Netherlands share similarities both social and political but that there are some important differences. In particular, the countries have experienced similar changes to immigration intake, and both are officially multicultural societies. The popular perception of immigration in the two countries remains quite different. Canada remains a country that actively seeks permanent migration, whereas the Netherlands is a country that is far more comfortable with the notion of temporary as opposed to permanent migrants. This in turn suggests that a comparison of engagement activity could yield valuable insight into the nature of participation on the part of majority and minority members of society. This is particularly the case since much of the literature suggests that minority participation will be lower than participation by majority members. We compare the participation rates using two independent surveys, one from Canada and the other from the Netherlands both of which were conducted at about the same time and both of which asked some similar questions.

3. Data, Measures and Methods

3.1 The Data

3.1.1 Canadian Data

Data for Canada are drawn from the second wave of the Equality Security and Community Survey (ESC).4 This is a large, national, and stratified random sample of the Canadian population who speak an official language (English or French) and are over eighteen years of age. Designed by a consortium of academics and administered by the Institute for Social Research (York University), it was conducted in 2002. We augment the sample by adding city size information from the 2001 Census.

The ESC survey is particularly suited for the investigation of issues of ethnicity and social capital for a number of reasons. First, the survey was administered to 5,654 respondents, 1452 of

4 The survey is also known as the Social Change and Well-being Survey (SCWB). 12 whom live in census tracts in Montreal, Toronto and Vancouver with four times the average number of visible minorities for that Census Metropolitan Area (CMA). This urban over-sample means that the total number of visible minorities available for meaningful analysis is substantial for a survey of this size. The survey has a large and varied set of questions on ethnic identity, ancestry and affiliation that far exceed the depth publicly available in other surveys, including the Canadian census. It also poses detailed questions related to forms of participation, and formal and informal kinds of social interaction which are directly comparable to data drawn in the Dutch sample.

3.1.2 Dutch Data

Data for the Netherlands come from the SPVA, the Social Position and Use of Facilities by Immigrants survey for the year 2002. The survey was carried out by the Institute for Sociological and Economic Research (ISEO) at Erasmus University in Rotterdam, the Dutch Social and Cultural Planning Office (SCP) and the Dutch Interdisciplinary Demographic Institute (NIDI) and aims to describe and analyse the socioeconomic and cultural integration of the four largest ethnic minorities in the Netherlands: the Surinamese, the Antilleans, Turks and Moroccans. The fieldwork was carried out from February 2002 until April 2003 – a period in which the societal discussion on immigration and integration in the Netherlands was at its peak due to the growth of the LPF (Lijst Pim Fortuyn, populist political party) and the aftermath of the September 11th attack.

Comparing the results of the survey to other sources shows that there is a slight over- representation of the elderly and an under-representation of those who live alone. Respondents are selected if they or their parents are born in one of the four groups listed above. To diminish problems with the survey, the interviewer and respondent are matched by ethnic group and the questionnaires were available in Dutch, Turkish and Arabic. Given the geographical concentration of the ethnic minority population, the SPVA consists of random samples of the population in thirteen cities, including the four largest, in the Netherlands.5

3.1.3 Limitations of the data

The two surveys ask a number of similar questions which we use to measure and compare rates of participation. However there are several limitations both because the overall goals of the surveys were different and because some concepts differ in the two countries. Prime among these is

5 For more information on the Dutch survey see Groeneveld & Weijers-Martens, (2003). 13 the level at which recreational participation can be compared. While the Dutch survey allows us to ask questions about the nature of the participation, such as whether the organization membership is primarily ‘minority’ based, information on the Canadian survey is much more limited. Thus while we can ask questions about whether people participate and the determinants of such participation, we cannot compare the degree to which these are either bridging or bonding activities.

Differences in minority populations also make it difficult to make group by group comparisons. The dominant minority groups in the Netherlands are Surinamese, the Antilleans, Turks and Moroccans, while the dominant minority groups in Canada are from Europe (i.e.: Italian and German), as well as from Asia (India and ) and the Caribbean (Jamaica). In order to compare regression results we are limited to looking at the population of all minorities or limiting our analysis to looking only at non-European minorities as compared to the majority population. Finally, differences in the education systems limit us to very coarse education comparison.

3.2 Derivation of the Variables

3.2.1 The Dependent Variable

The variable measuring recreational participation is calculated by respondents’ answers to questions dealing with participation in recreational activities. In Canada, respondents are asked the following: “How many recreational groups, such as sports leagues or clubs, music or hobby clubs, or exercise classes are you involved in?” The variable is made dichotomous: respondents participate in at least one activity or in none at all. In the Netherlands, respondents are asked in separate questions whether they are “a member of a sports club”, “a singing club”, and “a hobby club”. A variable is created from the three, and is then dichotomised.

3.2.2 Derivation of the independent variables

Our independent variables come in two varieties: a contextual variable which defines the size of the respondent’s community and individual-level variables. Community size is important for two reasons: first, communities may provide differential access to forms of recreational participation, in terms of availability, quality, and quantity of activities. Second, respondents in larger communities may exhibit different attitudes and behaviours than individuals in smaller communities (see Guest and Wierzbicki, 1999; Aizlewood and Pendakur, 2005). In the Canadian sample, each respondent is 14 linked on a case basis to census subdivision (CSD) population data from the 2001 census. In this way, we determine the natural log of the total population of the CSD in which a respondent lives (size of the community). In the Dutch sample, the log of the city population is added to each case

Individual level variables are divided into two types. The first type includes the basic sociodemographic markers that exist independently of ethnocultural ancestry but are implicated in discussions of participation. This group includes age, sex, educational attainment, marital status, employment status (employed/not working), and whether the respondent has children living in the household. The second group includes variables that refer to ethnic and cultural attributes. This group includes ethnic ancestry (minority/majority group), household language (official/non-official), religious affiliation, frequency of attendance at religious services, and for immigrants only, the age at which the respondent arrived in the country. The appendix contains details of how these variables were derived from the Canadian and Dutch datasets.

3.3 Analysis

The data are explored in several stages. In the first stage, we present simple cross-tabulations that examine the relationship between participation and ethnic and cultural attributes in Canada and the Netherlands. From these results, we identify basic patterns in the data and prepare the analysis for the second, more advanced, analytical stage. In the advanced stage, we use four phased, logistic regression models to evaluate the determinants of participation in the two countries. In the base-line phase, standard individual demographic control variables and the log of community size are tested for their effect on the odds of participation. Added at the second phase is a set of immigrant identifiers which define the age at which the respondent arrived in Canada or the Netherlands.6 In the third step, we add a variable that measures degree of religiosity. In the fourth and final phase, we add specific ethnocultural elements. We explore at each step the impact of each independent variable on the likelihood of participation in recreational activities.

6 We tested both age at immigration and period of immigration as well as a model with both variables. Age at immigration was chosen because we felt it provides a better measure of socialization. We also tested models with age at immigration interacted with period of immigration and age at immigration interacted with age. While the latter model did provide some additional insight about the nature of participation, missing categories between the two countries meant that comparison was impossible. We thus chose to concentrate on main effects models. 15

4. Findings

4.1 Basic findings

The descriptive statistics in the following Tables 1 through 4 provide an initial picture of the differences in recreational participation between the Netherlands and Canada for three key variables related to ethnic minority status. In Table 1, we see that overall, the participation rate in sports, singing and hobby organisations was far lower in the Netherlands than in Canada (20% vs. 48%). In Canada, there was little difference between majority and minority populations (48% vs. 47%). Among the majority population in the Netherlands, we found that 36% participated in these activities, as compared to only 14% of the minority population.

Table 1: Recreational Participation Among Majority and Minority Populations Canada The Netherlands Majority Minority Total Majority Minority Total Total 3,401 2,240 5,641 1,427 3,960 5,441 Yes 48% 47% 48% 36% 14% 20% No 52% 53% 52% 64% 86% 80%

Table 2 shows differences in recreational participation rates by for the two countries. In Canada, the breakdown shows little variation among the religious groups. Muslims and members of ‘other ’ have the lowest rates of participation (37% and 44% respectively), while those who report no religion have the highest (50%). In the Netherlands, a similar pattern is seen; those who identify themselves as Muslim and ‘other religion’ have the lowest recreational participation rates (10% and 9% respectively) and those who report no religion have the highest (29%).

Table 2: Recreational Participation Among Religious Denominations Other Other None Christian Total Eastern Religion Canada Total 1,192 3,896 132 211 210 5,641 Yes 50% 48% 37% 44% 47% 48% No 50% 52% 63% 56% 53% 52% Netherlands Total 2,025 1,253 1,814 275 74 5,441 Yes 29% 20% 10% 19% 9% 20% No 71% 80% 90% 81% 91% 80%

As shown in Table 3, speaking a non-official language at home has a clear negative relationship with recreational participation in both countries. This finding is expected, since previous 16 studies have identified inadequate language skill as a well-recognized barrier to recreational participation (see Stodolska, 2000). In Canada, only 37% of those who do not speak an official language at home participate in recreational activities as compared to 49% of those who speak English or French. In the Netherlands, only 13% of non-Dutch speakers participate, as compared to 87% of those who only speak Dutch in the home.

Table 3 Recreational Participation Among Official and Non-official Language Groups Canada The Netherlands Non- Non- Official Total Official Total official official Total 5,010 631 5,641 1,660 3,781 5,441 Yes 49% 37% 48% 35% 13% 20% No 51% 63% 52% 65% 87% 80%

Table 4 shows the breakdowns for immigrants and those born in Canada or the Netherlands. In Canada, immigrants participate in recreational activities at a lower rate than native-born Canadians (41% vs. 49%). In the Netherlands, differences between immigrants and non-immigrants are much more pronounced: only 13% of immigrants participate in these activities, as compared to 35% of the native-born Dutch.

Table 4 Recreational Participation Among Non-immigrants, Immigrants and Immigrants’ Age at Migration Immigrants’ Age at Migration Non- 14 and Over Immigrant Immigrant 15-20 21-30 Total under 30

Canada Total 4,700 941 466 98 165 212 5,641 Yes 49% 41% 38% 46% 50% 40% 48% No 51% 59% 62% 54% 50% 60% 52% Netherlands Total 1617 3824 717 855 1460 792 5441 Yes 35% 13% 23% 15% 11% 6% 20% No 65% 87% 77% 85% 89% 94% 80%

Table 4 also shows the breakdowns among immigrants on the basis of their age at the time of their arrival in Canada and the Netherlands. The findings suggest that for Canada, there is no clear pattern on the basis of age at migration. While one might assume that early arrival would bring about recreational participation rates at rates more closely matching those of non-immigrants, this does not appear to be the case. The age group most closely approximating the general population are those who arrived in Canada in their twenties. The group arriving before the age of fourteen are the least likely to participate. In the Netherlands, the expected pattern is seen: those arriving at younger ages 17 participate in recreational activities at higher rates than those who arrive when they are older. Only 6% of those who arrive in the Netherlands after the age of 30 participate in recreational activities.

4.2 Logistic regression models We now examine the phased impact of selected sociodemographic factors on the likelihood of individuals to participate in recreational activities. The tables present two forms of output from four logistic regression models: the odds ratios for each independent variable and their associated significance tests, and the results from independent samples t-tests that measure whether the impact of the independent variables is significantly different between Canada and the Netherlands.

4.2.1 The base-line model

The right side of Table 5 provides results for a baseline model. This model shows the effects of city size and sociodemographic factors on the probability of participating in recreational organisations/activities.7 The last column of the model identifies whether or not the coefficients for the two countries are significantly different from each other. 8 A dagger (†) identifies variables in which the t test value for the difference between the two countries is greater than 1.96.

The variable that measures participation in relation to the respondents’ city size shows for both countries that the size of the city has a significant negative effect on the participation rate. In both Canada and the Netherlands, the bigger the city in which a respondent lives, the lower the likelihood that that individual will participate in a recreational activity.

Males in the Netherlands are 29 per cent more likely to participate than females whereas no difference is measured in Canada. In both Canada and the Netherlands, age is inversely correlated with participation – in general, the older the individual, the less likely the person is to participate in a recreational activity. Further, the coefficients for age are not significantly different between the two countries.

7 In all cases, robust standard errors are used. 8 We determine if there is a significant difference between the two variables by calculating the t value for independent samples: coef − coef t = Canada sNetherland 2 2 SECanada + SE sNetherland a t value greater than 1.96 is taken as significant 18

In this basic model, individuals who are married in the Netherlands have a significantly lower probability (-17 per cent) of engaging in recreational participation. No significant effect is measured for Canada. Having children has a clear and significant negative effect on the probability of participation in the Netherlands, but no significant effect is measured for Canada.

Looking at the impact of education and employment on recreational activities, we find strong effects on the probability of being involved in recreational organisations in both countries. Higher levels of schooling and being employed increases the probability of participation in both countries. Thus, people with a university degree are 2.7 times more likely to participate in Canada and almost 4 times more likely to participate in the Netherlands. Looking at t values, we can see that only four coefficients are significantly different from each other (being male as compared to female, being married, having children and being employed). The other coefficients (city size, age and education) have about the same effect in the two countries.

Table 5, The probability of participating in formal recreational activity in The Netherlands and Canada. Models 1 and 2 Model 1 Model 2 Canada Netherlands T-test Canada Netherlands T-test Log of city size -0,03** -0,08*** 0,00 -0,05** Sex Males (Females) 0,03 0,29*** † 0,05 0,30*** † Age 25-34 (18-25) -0,29*** -0,24* -0,29*** -0,20 35-44 -0,35*** -0,25* -0,38*** -0,14 45-54 -0,44*** -0,33** -0,48*** -0,20 † 55-64 -0,45*** -0,30* -0,49*** -0,16 † 65+ -0,34*** 0,05 -0,42*** 0,12 † Marital status Married (not married) 0,11 -0,17** † 0,13* -0,17** † Children With (Without) -0,09 -0,45*** † -0,06 -0,30*** † Socio-economic Secondary (primary) 1,24*** 1,50*** 1,24*** 1,02*** University 2,73*** 3,94*** 2,87*** 2,52*** Employed (not emp.) 0,24*** 0,79*** † 0,18** 0,68*** † Age at migration 0-14 (native-born) -0,54*** -0,22** † 15-19 -0,26 -0,44*** 20-29 -0,06 -0,60*** † 30+ -0,29** -0,72 † R2 0,04 0,11 0,05 0,13 Significance: *** is 0.01, ** is 0.05 and * is 0.1 “†” indicates that a significant difference exists between the two countries for the t-test (>1.96).

19

4.2.2 The impact of migration factors

In the second model (right-hand side of Table 5) we add four dummy variables which identify the age at immigration. In this case, the comparison group is the native-born population. For the Netherlands we see a significant and increasing negative effect on participation when individuals have arrived at a latter stage in life. For Canada we see negative and significant effects for those who arrived as children and those who arrived when they were 30 or more years of age. Except for the case of those who immigrated when they were aged 15-19, the coefficients for the two countries are significantly different from each other.

For Canada, variables included in the previous model (age, education and employment status) maintain the same direction, magnitude and significance. This is not the case for the Netherlands. The addition of immigrant markers substantially reduces the impact of age in the model to the point where there are no significant differences across age groups. Further the impact of education is strongly reduced. Thus, whereas in Model 1 having a university degree resulted in a probability of participating of about four times that of people with low levels of schooling, this is now reduced to about 2.5 times. Comparing between the two countries, it is apparent that the education effects are about the same in both countries.

4.2.3 The impact of religiosity

Table 6 shows results for two additional models which add complexity to the analysis. The left-hand side of Table 6 provides results for Model 3, which includes all the previous variables and adds religious attendance. As can be seen, the explanatory power of this model is about the same as that found for Model 2 (shown in Table 5). With relatively few exceptions, the strength and significance of variables repeated in this model are about the same as well. Thus, age variables are similar in strength and direction for both Canada and the Netherlands. Being married has a weakly significant positive effect in Canada (.12) and a weakly significant negative effect in the Netherlands (-.17). Having children has no impact on participating in Canada, but a strongly negative impact in the Netherlands (-.28). Schooling has about the same impact as was seen in the previous model for both countries. The odds of participating in a recreational activity if employed remain about the same with no significant difference between the two coefficients. Age at immigration also has about the same impact as was seen in Model 2.

20

Attending a religious service has a positive but uneven effect in Canada as compared to the Netherlands, where religious attendance has no impact or a negative effect. People who attend religious services are more likely to participate in Canada compared to those who do not attend. The impact of attendance ranges from 18 per cent more likely for those who attend a few times a year or weekly to 37 per cent more likely for those who attend a few times a month. This does not hold true for the Netherlands, where people who attend are either less likely or as likely to participate as those who do not attend. Save for those who participate only a few times a year, the coefficients for each country are significantly different from each other.9

Table 6, The probability of participating in formal recreational activity in The Netherlands and Canada. Model 3 Model 4 Canada Netherlands T-test Canada Netherlands T-test Log of city size -0,00 -0,05** -0,02 -0,05** † Sex Males (Females) 0,06 0,33*** † 0,07 0,38*** † Age 25-34 (18-25) -0,29*** -0,22 -0,36*** -0,28** 35-44 -0,38*** -0,17 -0,42*** -0,26* 45-54 -0,49*** -0,23 -0,52*** -0,35** 55-64 -0,50*** -0,18 † -0,52*** -0,33** 65+ -0,43*** 0,10 † -0,39*** 0,08 † Marital status Married (not married) 0,12* -0,17** † 0,19** -0,12 † Children With (Without) -0,08 -0,28*** -0,09 -0,27*** Socio-economic Secondary (primary) 1,23*** 0,96*** 1,31*** 0,87*** University 2,88*** 2,37*** 2,87*** 2,10*** Employed (not emp.) 0,18** 0,64*** 0,22** 0,55*** Age at migration 0-14 (native) 0,55*** -0,19* † -0,33** -0,02 15-19 -0,28 -0,40*** -0,38 -0,26* 20-29 -0,07 -0,56*** 0,11 -0,44*** † 30+ -0,29** -0,69*** -0,18 -0,61*** † Religious attendance Few times a year (never) 0,18** -0,02 0,37*** -0,02 Few times a month 0,37*** -0,18 † 0,46*** -0,12 † Weekly 0,18** -0,45*** † 0,37** -0,38*** † Religion Christian -0,19* 0,11 Muslim -0,44* -0,24** Other Eastern -0,05 0,28 Other -0,29 -0,45 Ethnicity Minority Home Lang(Maj.) -0,55*** -0,23* † Minority Ethnicity (Majority) -0,08 -0,02 R2 0,05 0,13 0,05 0,13 Significance: *** is 0.01, ** is 0.05 and * is 0.1 “†” indicates that a significant difference exists between the two countries for the t-test (>1.96).

9 We tested a model which included an interaction between religion and sex. This model did not change the results very much and in fact, the two variables appear to be independent of each other. 21

4.2.4 The impact of ethnocultural and religious characteristics

The fourth model (right-hand side of Table 6) adds markers of ethnic affiliation. These include religious affiliation, whether the respondent uses a minority home language and whether the respondent is a member of a minority community. Because of differences in the types of minority groups between the two countries we drop respondents who report only Aboriginal or European origins and restrict the Canadian sample to respondents who are or majority (reporting either British, French of Canadian origin).10

Adding these variables does not change the direction, magnitude or significance to any great degree of any of the previously included variables. One important exception is that of city size. In prior models, city size tended to be negative in the Netherlands (at about 5 percent, for every increment of the log of city size) and small and insignificant in Canada. However, there was no significant difference between the two coefficients for the two countries. With the addition of the ethnic marker variables the variances for city size are tightened up to the degree where there is a significant difference between the two countries. Thus, the impact of city size is small and insignificant in Canada and quite strongly positive in the Netherlands (5 per cent for every value additional log value).

Looking first at the religion variables, we see that for both Canada and the Netherlands, religious affiliation has a negative or insignificant effect. Thus, being Muslim results in lower levels of recreational participation in the Netherlands -24 per cent) and a negative, but weakly significant level of participation in Canada (-44%) as compared to not stating a religion. Being Christian in Canada results in weakly significant lower levels of participation, but not in the Netherlands, and in fact, the two coefficients are not significantly different from each other. Coefficients for the two other religion categories are not significant.

Speaking a minority language at home has a significant and negative impact in both Canada and the Netherlands (-.55 per cent and -.23 per cent respectively). However, the addition of being a member of a minority ethnic group does not have a significant impact in either country, suggesting

10 An analysis conducted for the entire sample yields similar results in that coefficients tend to be about the same magnitude and the same direction as well as the same level of significance. However, dropping European-origin Canadian minorities more closely matches the Dutch minority population so it makes sense to do so for purposes of comparison. 22 that the other cultural variables (religion and minority home language are picking up the explanatory power of minority status).

Overall the analysis suggests that the strongest predictors in Canada are age, education, employment status and religious attendance. Immigrants, except for those who arrived when they were relatively young, do not show significantly different propensities to participate. Where there are differences by religion, they are only weakly significant. Speaking a minority language at home however, has a strong, negative effect in Canada. In the Netherlands, while similar results are found for socio-economic variables (age, education and employment), the differences by minority markers tend to be more pronounced. For immigrants, higher ages at entry, are correlated with lower rates of participation. Those claiming Muslim as religion are also less likely to participate. However, the effect of speaking a minority language at home is not nearly as pronounced.

5. Discussion and Conclusions

Participation in community events and activities is an important part of community interaction. Prior research has demonstrated that such participation is important for overall life satisfaction and can lead to other benefits, such as increased interaction across diverse groups and an increased sense of belonging to one’s community. The purpose of our study is to analyse the rates at which immigrants and ethnocultural minorities in Canada and the Netherlands participate in community activities such as sports, music or hobby activities and assess the determinants of such participation. We examined first the rates of immigrant and minority/majority community engagement as measured by recreational activity in the Netherlands and Canada. Second, we considered the role that language, religion, urban living, age and education and other factors play in the likelihood of recreational participation. Third, we assessed the importance of ethnocultural characteristics in explaining the probability of participation as compared to sociodemographic characteristics. Our study is limited because while we can examine recreational participation overall, we are unable to break it into either bridging (between group) or bonding (within group) recreational activities. We are, however, able to look at differences between Canada and the Netherlands and assess the degree to which the likelihoods of participation are significantly different from each other. Analysis of basic descriptives suggested that the overall rate of recreational participation is higher in Canada than in the Netherlands. The results suggest that ethnic minority status, immigrant status, belonging to a minority religion, and speaking a non-official language all exert negative influences on 23 the rate at which individuals participate in recreational activities. These findings appear to confirm the findings of other studies that identify these elements as negative correlates with rates of participation, in particular those of Stodolska (1998; 2000), Stodolska and Alexandris (2004) and Rublee and Shaw (1991) who found that ethnic minorities and immigrants face particularly acute constraints due to language, lack of knowledge, lack of time, and fatigue. Further, there appear to be stronger correlations with minority status in the Netherlands than is the case for Canada; minorities in that country exhibit lower levels of participation than those in Canada. The logistic regressions confirm this to a degree, but also point to the overriding importance of age, schooling and employment status. The findings around education and employment status tend to support some elements of the opportunities hypothesis, at least in the Netherlands. This hypothesis states that as costs increase, opportunities for marginalized groups decline. We find that in the Netherlands, individuals with lower levels of schooling and those who are unemployed are less likely to participate.

There are some other marked differences between the two countries. The size of the city in which a respondent lives has a strong and negative impact in the Netherlands across all four models, but is insignificant in Canada in all but the first model. Men are far more likely to engage in recreational activity as compared to women in the Netherlands, but are no more likely in Canada. Being married has a positive impact in Canada, but an insignificant impact in the Netherlands. Immigrant status and age at immigration are generally not important factors in determining participation in Canada, but are negatively correlated in the Netherlands.

Some similarities between the countries are also seen. In both countries, when controlling for socio-economic and ethno-cultural characteristics, Muslims exhibit lower rates of participation. Lower rates of participation are also seen in those who speak a minority language at home. These findings tend to support those of Alesina and Ferrara (1999), Verma and Darby (1994), de Knop, et al. (1994) and Duyvendkak, et al. (1998), who found that religious affiliation exerted a negative influence on rates of participation in general. We find, however, no additional effect of minority status beyond language and religion; minorities who speak the majority language at home and who do not have strong religious ties exhibit roughly the same participation attributes as do the native-born in both countries. Overall, the strongest determinants of participation in recreational activity are socio- demographic in nature – age, education and employment status. Not only is the impact of these variables significant in both countries, but the effects are of roughly the same magnitude. Indeed, the differences in the Netherlands between majority and minority respondents largely ‘wash out’ when socio-economic characteristics are added to the analysis. 24

Given that the participation rate of minorities in the Netherlands is lower than in Canada – but that much of this can be explained by socio-economic status (education and employment) – our findings suggest that on average, the socio-economic attributes of immigrants in the Netherlands is lower than in Canada. Thus, despite differences in migration policy and intake patterns, minorities participate in a fashion similar to the native-born population. One exception to this is the impact of speaking a minority language at home, which has a much stronger negative impact in Canada than in the Netherlands. Such a result may be a product of lower levels of language training in Canada as compared the Netherlands, or may be an outcome of the funding programs available to Dutch migrants for minority participatory activities (Van Heelsum, 2004).

Overall, our findings suggest that issues around education and employment status may be the more dominant determinants of participation rather than those of ethnocultural minority status, immigrant status, or religious affiliation in particular. The common factor across the two countries – despite some differences in terms of age at immigration or language preference – is the clear dominance of education and employment status. Thus, the achieved characteristics of individuals have the potential to far outweigh their ascribed characteristics in the area of recreational participation. While individuals cannot readily change their ethnic origin, education levels can be increased, and socioeconomic status can be raised. This finding is important and positive, since recreational participation in multicultural societies has the potential to increase the engagement and interaction of like and unlike groups. Such participation can lead to other social benefits, such as increased understanding across diverse groups and an increased sense of belonging to one’s community.

25

6. References

Aizlewood, A. and R. Pendakur. (2005) “Ethnicity and Social Capital in Canada”, Canadian Ethnic Studies. Vol. XXXVII, No. 2, 2005.

Alesina, A., and E. La Ferrara. (1999) “Participation in Heterogeneous Communities”, The Quarterly Journal of Economics 115(3): 847-904.

Brehm, J. and W. Rahn. (1997). “Individual-level Evidence for the Causes and Consequences of Social Capital,” American Journal of Political Science 41(3): 999-1024.

Breton, R. (1997). “Social Participation and Social Capital” Paper presented at the Second National Metropolis Conference, Montreal, November.

Breton, R. (2003). “Social Capital and the Civic Participation of Immigrants and Members of Ethnocultural Groups” Paper presented at the Conference on The Opportunities and Challenges of Diversity: A Role for Social Capital? Montreal: November.

Coakley, J. (2001). Sport in Society: Issues and Controversies (7th Edition). -Toronto: McGraw Hill.

Dagevos, J. (2001). Perspectief op Integratie. Over de Sociaal-Culturele en Structurele Integratie van Etnische Minderheden in Nederland. Den Haag: WRR.

De Haan, J. & Breedveld, K. (2000). Trends en determinanten in de sport. Erste resultaten uit het AVO 1999. Den Haag: SCP Werkdokument.

De Hart, et al. (2002). Zekere banden, Sociala cohesie, leefbaarheid en veiligheid, Den Haag: SCP.

De Knop, P., Wylleman, P., Theeboom, De Martelaer, K., Van Puymbroek, L. & Wittock, H. (1994). Youth-friendly sport clubs. Developing an effective youth sport policy. , Belgium: VUB Press.

Distelbrink, M. & Veld, T. (2001). “Sport”, in Minderhedenmonitor 2000, Various authors, Instituut voor Sociologisch-Economisch Onderzoek (ISEO), Rotterdam and Centrum voor Onderzoek en statistiek (COS), Gemeente Rotterdam.

Doomernik, J. (2005), “The State of Multiculturalism in the Netherlands” in Canadian Diversity, Vol. 4:1.

Duyvendak, J.W. et al., (1998) Integratie door sport? Een ondezoek naar gemengde en ongemengde sportbeoefening van allochtonen en autochtonen. Rotterdam: Gemeente Rotterdam.

Fennema, M. et al., (2000), Sociaal Kapitaal en Politieke Participatie van Etnische Minderheden. Amsterdam: IMES.

Floyd, M.F. (1998) “Getting Beyond Marginality and Ethnicity: The Challenge for Race and Ethnic Studies in Leisure Research” Journal of Leisure Research. Q1 Vol. 30: Issue 1.

Forrest, R. and Kearns, A. (2001). “Social Cohesion, Social Capital and the Neighbourhood,” Urban Studies 38 (12): pp.2125-43. 26

Fukuyama, F. (1995). Trust: The Social Virtues and the Creation of Prosperity. New York: Free Press.

Gordon, M. (1964). Assimilation in American Life. New York: Oxford University Press.

Groeneveld, S. & Weijers-Martens, Y. (2003), Minderheden in beeld, SPVA-02, Instituut voor Sociologisch-Economisch Onderzoek (ISEO), Rotterdam.

Heelsum van, A. (2004). Migrantenorganisaties in Nederland, Deel 2: Functioneren van de organisaties. Utrecht: Forum.

Helliwell, J. and R. Putnam. (1999). Education and Social Capital. Working Paper No. 7121. National Bureau of Economic Research, Cambridge, MA.

Helly, D. (1997). “Voluntary and Social Participation by People of Immigrant Origin: Overview of Canadian Research”. Paper presented to the Immigrants and Civic Participation: Contemporary Policy and Research Issues session of the Second National Metropolis Conference. Montréal, November.

Henderson, K.A., & Ainsworth, B.E. (2001). “Social support, constraints, and time for physical activity involvement: Perspectives of American Indian women from the Cultural Activity Participation Study”. In K.J. Daly (Ed.), Minding the time in family experience, Vol. 3 (pp. 383-402). : Elsevier Science Ltd.

Hutchison, R. (1988). A Critique of race, Ethnicity, and Social Class in Recent Leisure-Recreation Research. Journal of Leisure Research, vol. 20 p.10-30.

Hutchison, R. (1987). Ethnicity and Urban Recreation: Whites, Blacks and Hispanics in Chicago’s Public Parks. Journal of Leisure Research, Vol. 19, p. 205-222.

Jackson, E.L. and E. Dunn, (1988). "Integrating Ceasing Participation with Other Aspects of Leisure Behavior". Journal of Leisure Research. Vol. 20, No. 1, pp. 31-45

Johnson, C.Y., P.M. Horan and W. Pepper (1997). “Race, Rural Residence, and Wildland Visitation Examining the Influence of Sociocultural Meaning”. Rural Sociology Vol. 62 p. 89-110.

Juniu, S. “The impact of immigration: Leisure experience in the lives of South American immigrants”. Journal of Leisure Research, 32(3), 358 - 381.

G. Karlis and D. Dawson “Ethnic Maintenance and Recreation: A Case Study” Journal of Applied Recreation Research. Volume 15, Number 2

Lagendijk, E. & Van der Gugten, M. (1995). Sport en allochtonen: feiten, ontwikkelingen en beleid (1986-1995). Amsterdam: Van Dijk, Van Soomeren en Partners BV.

Ley, David, and A. Germain. (2000). “Immigration and the Changing Social Geography of Large Canadian Cities.” Plan Canada Vol. 40, No. 4

Lindsay J.J. and R.A. Ogle (1972). Socioeconomic Patterns of Outdoor Recreation Use Near Urban Areas. Journal of Leisure Research. Vol. 4, p. 19-24. 27

Martinez, S.D. & Vreeswijk, A. (2002). ‘Ruimtelijke concentratie van allochtonen’ in De toekomst in meervoud, Perspectief op multicultureel Nederland, (Eds.) J. Veenman, Assen: Koninklijke Van Gorcum.

Nie, N. H., Junn, J., & Stehlik-Barry, K. (1996). Education and democratic citizenship in America. Chicago: University of Illinois Press.

Pendakur, R. (2000). Immigrants and the labour force: policy regulation and impact. Montreal: McGill-Queen’s University Press.

Penninx, R. & Schrover, M. (2001) Bastion of bindmiddel? Organisaties van immigranten in historisch perspectief. Amsterdam: IMES.

Putnam, R. (1993). Making Democracy Work: Civic Traditions in Modern Italy. Princeton: Princeton University Press.

Putnam, R. (1995). “Bowling Alone: America’s Declining Social Capital,” Journal of Democracy 6:1 (January): 65-78.

Rublee C.B. and S.M. Shaw (1991). “Constraints on the Leisure and Community Participation of Immigrant Women: Implications for Social Integration” Loisir et Société 14, p. 133-150.

SCP (Sociaal Cultureel Planbureau), (2000). Sociaal en cultureel rapport 2000, Den Haag: Sociaal en Cultureel Planbureau.

Stodolska, M. (1998). “Assimilation and Leisure Constraints: Dynamics of Constraints on Leisure in Immigrant Populations” Journal of Leisure Research Q4 Vol. 30, Issue 4.

Stodolska, M (2000). Changes in leisure participation patterns after immigration. Leisure Sciences, 22, 39-63.

Stodolska, M., & Alexandris, K. (2004). The role of recreational sport in the adaptation of first generation immigrants in the United States. Journal of Leisure Research, 36, 379-413.

Tirone, S., & Shaw, S. M. (1997). “At the center of their lives: Indo Canadian women, their families and leisure”. Journal of Leisure Research, 29, 225-244.

Uunk, W. (2002). Concentratie en achterstand, Over de samenhang tussen etnische concentratie en de sociaal-economische positie onder allochtonen en autochtonen, ISEO: Koninklijke Van Gorcum.

Walter, T.O., B. Brown and E. G. Grabb (1991). "Ethnic Identity and Sports Participation: A Comparative Analysis of West Indian and Italian Soccer Clubs in Metropolitan Toronto," Canadian Ethnic Studies, Vol. 23, No. 1: 85-96.

Washburne, R.E. (1978). Black Under-Participation in Wildland Recreation: Alternative Explanations. Leisure Sciences, Vol. 1, p. 175-189.

Woodard, M.D. (1988). Class, Regionality and Leisure Among Urban Black Americans: The Post- Civil Rights Era. Journal of Leisure Research Vol. 20, p. 87-105. 28

Yancey, W.& Snell, J. (1976). Parks as aspects of leisure in the inner city: An exploratory investigation. Research in the Parks - Transactions of National Park Centennial Program Symposium. American Association for the Advancement of Science. USDI Symposium Series. Number 1, 146-168.

29

7. Appendix

This appendix describes the variables used in the analyses and their basic descriptive statistics (see also Tables A1 and A2). Variables for Canada are obtained from the 2002 Social Capital and Well- being survey or from the 2001 Canadian Census linked through forward sortation postal codes and Census subdivisions. For the Netherlands the 2002 SPVA survey (Social Position and Use of Facilities by Immigrants) is used. Most variables have been recoded, combined or reconstructed for the purposes of the analyses.

Recreational participation

Canada: dichotomous variable based on the question “How many recreational groups, such as sports leagues or clubs, music or hobby clubs, or exercise classes are you involved in?” Values are ‘yes’ (respondent is involved in at least one recreational group), and ‘no’ (respondent is not involved in any recreational group). The latter forms the reference category. Netherlands: dichotomous variable computed from the questions “Are you a member of a sports club?”, “Are you a member of a singing/music/acting club?” and “Are you a member of a hobby club?” The new variable is then dichotomised. Values are ‘yes’ (respondent is involved in at least one recreational group), and ‘no’ (respondent is not involved in any recreational group). The latter forms the reference category.

Community size

Canada: the natural log of the population of the respondent’s municipality of residence, as derived from the 2001 Canadian Census. Netherlands: the natural log of the population of the respondent’s municipality of residence, as derived from Statline, Statistics Netherlands.

Age

Canada and the Netherlands: dummy variable computed from the question “In what year were you born?”. Values are categorised into six values ‘under 25’, ‘25-34’, ‘35-44’, ‘45-54’, ‘55-64’ and ‘65 and older’, where ‘under 25’ forms the reference category.

Sex

Canada and the Netherlands: dichotomous variable in which the respondent is ‘male’ and the reference category is ‘female’.

Educational attainment

Canada: dummy variable based on the question “What is the highest level of education that you have completed?” Netherlands: dummy variable based on the question “What is the highest level of education that you have completed in the Netherlands?” and/or “What is the highest level of education that you have completed outside the Netherlands?” Values are ‘post-secondary complete’, ‘some post-secondary’, and ‘secondary’ and ‘none/primary’. The latter forms the reference category.

30

Marital status

Canada: dichotomous variable based on the question “Are you presently married, living with a partner, separated, divorced, widowed, or have you never been married?”. Netherlands: dichotomous variable based on the questions “What is your civil status?” and “ Do live together with your partner/husband/wife?” Values are ‘married/has partner/common law’ and ‘single/alone’. The latter forms the reference category.

Respondent has children

Canada: dichotomous variable based on the questions “How many children do you have, including any no longer living with you?” and “Does this child/How many of these children are currently living with you for four or more days a week?”. Netherlands: dichotomous variable based on the questions “Do you have children both in the Netherlands and abroad? and How many children do live in your home? Values are ‘has child(ren) living at home’ and ‘no child(ren) living at home’. The latter forms the reference category.

Employment status

Canada: dichotomous variable based on the question “Are you currently self employed, working for pay, retired, unemployed, or looking for work, a student, a homemaker, or something else?” Netherlands: dichotomous variable based on the questions “Do you have paid work at the moment?” and “What kind of work do you at this moment?”. Values are 'not working' and 'working or self-employed'. The latter forms the reference category.

Age at migration

Canada: dummy variable based on computed value of the variable “In what year were you born?” subtracted from “In what year did you come to live in Canada?”. Values are 'Aged 14 and under', 'Aged 15-20' 'Aged 21-30' 'Aged over 30'. The reference category is ‘Canadian-born/non- immigrants’. Netherlands: dummy variable based on computed value of the variable “In what year were you born” subtracted from “In what year started you to live in the Netherlands?” Values are 'Aged 14 and under', 'Aged 15-20' 'Aged 21-30' 'Aged over 30'. The reference category is ‘Dutch-born/non-immigrants’.

Religiosity

Canada: dummy variable based on the question “How often do you attend religious services, not including weddings and funerals?” Values are ‘weekly’, ‘2-3 times per month’, ‘a few times a year’, and ‘never’. Netherlands: “How often do you attend a religious service?” Values are ‘one time or more per week’, ‘never’, ‘a couple of times a month’ and ‘a couple of times a year’. The reference category for both countries is a combination of respondents with no religion (N/A) and of those responding ‘never’.

Religion

Canada: dummy variable based on the question “Please tell me what your religion is, if you have one?” 31

Netherlands: dummy variable based on the questions “Do you count yourself to a certain religion, church or religion? and “Which belief is this?” Values are categorised as follows: ‘No religion’, ‘Christian’, ‘Islam’, ‘other eastern religion’, and ‘other religion’. ‘No religion’ forms the reference category.

Language spoken in the home

Canada: dichotomous variable based on the question “What language do you usually speak at home?” Values are ‘just official language (French, English)’ and ‘other language’. The latter forms the reference category. Netherlands: dichotomous variable based on the question “Do you speak Dutch with your partner?” Values are ‘Yes, often/always’, ‘Yes, sometimes’ and ‘No, never’. The latter two categories are pooled together and used as the reference category.

Ethnicity:

Canada: dichotomous variable computed from the questions “To what ethnic or cultural group do you belong?” and “In addition to being Canadian to what ethnic or cultural group did you, or your ancestors belong on first coming to this continent?”. Majority members are those who identified solely as British, French, Canadian or a combination of these groups. Minority members are those who identified as being a member of another ethnic group. Netherlands: dichotomous variable computed from the questions “In what country are you born?”, “In what country is your father born?” and “In what country is your mother born?” Values are ‘Dutch’ and ‘other’. The latter forms the reference category. Given that the survey oversamples minorities, the mean of .79 is substantially higher than the population mean. 32

Table A1: Canada - Descriptives

Variable N Mean SD Min Max Recreational participation 5641 0.48 0.50 0 1 Sex 5654 0.46 0.50 0 1 Employment Status 5498 0.62 0.48 0 1 Educational attainment 5550 1.76 1.03 0 3 Age in six categories 5533 2.56 1.52 0 5 Marital status 5621 0.59 0.49 0 1 Respondent has children 5639 0.67 0.47 0 1 Religion 5654 1.00 0.85 0 4 Religiosity 5654 1.22 1.18 0 3 Ethnicity 5654 0.40 0.49 0 1 Language 5654 0.11 0.32 0 1 Natural log of population 5654 11.28 2.42 2.30 14.72 Age at migration 5654 0.36 0.95 0 4

Table A2: The Netherlands - Descriptives

Variable N Mean SD Min Max Recreational participation 5441 0.20 0.40 0 1 Sex 5440 0.59 0.49 0 1 Employment Status 5441 0.58 0.49 0 1 Educational attainment 5365 1.72 0.92 0 3 Age in six categories 5441 2.35 1.35 0 5 Marital status 5414 0.54 0.50 0 1 Respondent has children 5441 0.59 0.49 0 1 Religion 5441 1.10 1.01 0 4 Religiosity 5441 0.99 1.17 0 3 Ethnicity 5343 0.79 0.41 0 1 Language 5441 0.69 0.46 0 1 Natural log of population 5439 12.62 1.21 0 13.51